Summary of Work: We study environmental effects on human reproduction, and the biology that underlies such effects. We published data this year showing that aging of human ova after ovulation and before fertilization may increase the risk of very early embryonic loss. This has been suggested in studies of laboratory animals, but there has been no previous evidence of such an effect among humans. As part of our effort to improve laboratory methods for studying human reproduction, we have shown that metabolic products of luteinizing hormone can be readily detected in women's urine even after years of frozen storage. This suggests that urinary measures of LH can be useful in epidemiologic studies, for example for identifying ovulatory cycles. The development of new methods is especially important for an emerging area such as reproductive epidemiology. In collaboration with colleagues in the Biostatistics Branch, we have published a method for analyzing fertility data that takes into account statistical dependency when several cycles of data are available from an individual woman. On more applied problems, we have reviewed factors that cause a baby to be born late, with subsequent health problems. Environmental toxins might contribute to post-term delivery, but little is known about causes of this problem. Also, we have shown a link between low fertility and the risk of preterm delivery. Since preterm delivery is a common cause of infant morbidity and mortality, this may provide some additional clues about its cause. Finally, we continue to develop the use of menstrual cycle data as an indicator of women's reproductive health, showing that a woman's smoking affects the amount of menstrual bleeding and menstrual pain.